🎭 LLMs Last Test: Memes 😂
Help us create the ultimate meme understanding dataset for AI research!
🚀 How it works:
Evaluate Memes
Analyze memes by evaluating descriptions and identifying their humor types, emotions conveyed, and cultural context. Help us understand what makes memes funny across different cultures!
Upload & Classify
Share your memes with detailed classifications including descriptions, humor analysis, emotional mapping, cultural reach, and temporal context.
Advance Research
Your contributions help create benchmark datasets for evaluating AI understanding of internet humor, cultural references, and multimodal reasoning in memes.
📚 About This Research
What is MemeQA?
MemeQA is a comprehensive crowd-sourced meme understanding dataset designed to evaluate Vision-Language Models (VLMs) in their comprehension of internet memes and human humor. It includes detailed cultural classifications, humor analysis, emotional mappings, and descriptions to create the most comprehensive meme research dataset available.
The dataset is being developed by researchers at THWS (Technische Hochschule Würzburg-Schweinfurt) and CAIRO (Center for Artificial Intelligence) within the NLP Team to address the growing need for AI systems that can understand not just what they see, but the cultural and humorous context behind visual content.
Research Goals:
- Evaluate VLM meme comprehension
- Benchmark humor understanding in AI
- Create diverse cultural context dataset
- Enable explainable meme analysis
- Study cross-cultural humor patterns
Features:
- 10+ classification dimensions per meme
- Emotional and humor type analysis
- Cultural reach & platform tracking
- Description analysis
- User performance analytics
NLP Team: MSc Rodolfo Cacacho • Prof. Dr. Ivan Yamshchikov
Website: CAIRO THWS • Contact: cairomemeqa@gmail.com